AppBespoke · Glasgow, working US afternoons
AI search optimization: get recommended by ChatGPT, Perplexity and Google AI
We find the buyer questions where AI recommends your competitors instead of you, fix the reasons in your site's code and content, and track the answers every week. Audit $750. Retainers from $1,500/month. No sales call.
- ChatGPT
- Perplexity
- Gemini
- Google AI Overviews
- Fixed prices
- Fixes in your code
- No sales call
Audit $750 fixed · report in 7 days · retainers from $1,500 a month
Free snapshot
10 buyer prompts for your category on ChatGPT and Perplexity: which ones name you, and which name a competitor. By the end of the next US business day. No call.
Get a free snapshotNo guaranteed mentions.What we won’t do, and whyWhat your buyers read before they reach you
Google says AI Overviews now has over 2.5 billion monthly users, and that AI Mode passed 1 billion within a year of launch. For a lot of buying questions, the AI answer is the first thing on the screen, above the links you worked to rank in.
When a buyer asks "what is the best tool for this", the answer names three or four companies. If yours is not one of them, you never find out it was asked. This work measures how often that happens and then works on it.
A buyer asks
“best project management software for small architecture firms”
The AI answer
3 brands named- 1.
Plankline. Built for small studios, with timesheets and client billing in one place.
- 2.
Drafthouse PM. Cheapest per seat, and the one reviewers mention for drawing sets.
- 3.
Studio Ledger. Strong on fee tracking, often recommended for firms under 20 people.
Your product
not mentionedThe category and brands above are made up. Your snapshot uses real prompts for your category and the real answers.
What does the audit actually find?
Five areas. The first two say where you are missing from the answers. The other three say why, which a tracking tool cannot tell you.
Prompts
50 prompts a buyer in your category would actually type: "best X for Y", "X vs Y", and the use-case questions. You agree the list before anything runs.
Example finding
31 of 50 prompts name at least one competitor. 6 name you, all on "X vs Y" questions you are already in.
Engines
Every prompt run on ChatGPT, Perplexity, Gemini and Google AI Overviews, with a screenshot of each answer. The engines disagree more than you would expect.
Example finding
Perplexity names you on 11 prompts. ChatGPT names you on 2 of the same 11.
Citation sources
Which pages each engine cites when it answers, for you and for up to 3 competitors. This is where their mentions come from.
Example finding
Two comparison articles and one G2 category page account for most competitor citations. None of them mention you.
Technical
AI crawler access in robots.txt and at your CDN, whether key pages render without JavaScript, structured data, llms.txt, speed on key templates, and internal links to the pages that sell.
Example finding
Pricing and feature pages render client-side only, so a crawler that does not run JavaScript sees an empty shell.
Content
Whether your key pages answer the questions buyers ask, in the form engines quote: direct answers, comparisons, specs, and a price someone can find.
Example finding
No page says who the product is for in one sentence, and pricing is behind a "contact sales" button.
What you get
A report, a plan and a video
A written report with a screenshot of every answer, a prioritised 90-day plan of at least 20 specific fixes, and a recorded walkthrough. No call needed.
$750 fixed, 7 days from the prompt list being agreed.
How does it work?
01
Free snapshot in 1 business day
10 buyer-style prompts written for your category, run on ChatGPT and Perplexity. Back comes which ones mention you and which mention competitors instead. Free, one per company.
02
Audit in 7 days
The prompt list is agreed first. Then 50 prompts, four engines, the source map, the technical and content checks, and a written report with a recorded walkthrough, 7 days later.
03
Retainer: pages, fixes, tracking
Pages written or rewritten in an answer-first shape, technical and structured-data fixes made in your code, and the prompts tracked every week across all four engines.
04
Monthly report and video
A written report every month and a recorded video: what the answers say now, what changed, AI referral traffic in your analytics, and what is next.
What we need from you
Nothing for the snapshot beyond the form. For the audit and the retainer, these three. None of it has to be ready to send the form.
- Access to the site, or a named developer to hand changes to
- Read access to your analytics
- Your list of competitors, if you have one. If not, we find them
What does the audit report look like?
Every fix in the 90-day plan carries the same four things: a priority, the area it sits in, what was found with a screenshot of the answer, and what to change. Here are three of the kinds that turn up.
AI crawlers are blocked at the CDN
- What was found
- robots.txt allows every bot, but the CDN blocks known AI crawlers at the edge, so none of them can read the site.
- What to change
- Allow GPTBot, PerplexityBot and Google-Extended in robots.txt, and turn off the CDN rule that returns 403 to them.
One roundup drives most competitor mentions
- What was found
- On 9 of 14 "best X" prompts, Perplexity cites the same third-party roundup. It lists four competitors and not you.
- What to change
- Ask the publisher to review the product for the next update, with a comparison page on your site they can cite.
Engines quote the wrong price
- What was found
- The pricing page is a table rendered in JavaScript with no text answer, so engines quote a review site's out-of-date price instead.
- What to change
- A short answer at the top of the page: who it is for, what it costs, and how it differs, followed by the detail.
These show the format, not a real client’s report. Yours covers your prompts, your competitors and your site, with a screenshot against each entry and a recorded walkthrough.
Ask about the auditTools, agencies, guarantees, and this
You are probably weighing some of these already. Here is what each one costs and what it changes.
| Option | What it costs | What it does |
|---|---|---|
| A tracking tool | $29-99 a month | Useful, and cheap. It tells you that you are missing from the answers. It does not tell you why, and it does not make the fixes. |
| A large SEO or AEO agency | $1,600-10,000 a month | The same goal as this. You usually get a strategy call first, then account managers and a writer network. Here the person who audits your site is the engineer who changes its code and writes the pages. |
| "Recommended or you don't pay" | Varies | Nobody controls what an AI model says. These offers usually lean on Reddit posting and review volume, which the platforms treat as manipulation. The risk sits with your brand, not theirs. |
| Doing nothing | Nothing up front | A real option. The answers keep being given either way, and the ones in your category currently name somebody. |
| This page | $750 audit, then from $1,500 a month | We improve what engines can find, read and quote on your site, and in sources you earn legitimately. The fixes go into your site's code as well as its pages. No promise of mentions, because nobody can make one. |
An engineer, not a content mill. The fixes go into your site’s code (crawler access, rendering, structured data, page structure) as well as the pages themselves.
What does it cost?
A free snapshot, a fixed-price audit, and two monthly retainers. Every number is here, and so is what would change it.
Free
by the end of the next US business day, no call
- 10 buyer-style prompts written for your category
- Run on ChatGPT and Perplexity
- Which prompts name you, and which name a competitor instead
- A one-page PDF or an email. One per company domain
$750
fixed, 7 days from the prompt list being agreed
- 50 buyer-intent prompts, the list agreed with you first
- 4 engines: ChatGPT, Perplexity, Gemini, Google AI Overviews
- Up to 3 competitors, and the sources engines cite for them
- Technical check: AI crawler access, rendering, structured data, llms.txt, speed, internal links
- Content check against the questions buyers ask
- Written report, a 90-day plan of at least 20 fixes, and a recorded walkthrough
Credited: the $750 comes off month one of a retainer started within 30 days.
Monthly retainer
3 months, then monthlyPages written, fixes made in your code, and the answers tracked every week. 3 months first, because AI answers move over weeks and a single month judges a signal that has not arrived yet.
Visibility
$1,500
a month
Prompt tracking, 4 engines, weekly100 prompts
Competitors tracked3
Answer-first pages written or rewritten4 a month
Technical and structured-data fixes in your site's codeUp to 4 h a month
Source-gap reportQuarterly
AI referral traffic in your analyticsMonthly
ReportingMonthly written report and recorded video
Term3 months, then monthly
Visibility Plus
$2,500
a month
Prompt tracking, 4 engines, weekly250 prompts
Competitors tracked5
Answer-first pages written or rewritten8 a month
Technical and structured-data fixes in your site's codeUp to 8 h a month
Source-gap reportMonthly
AI referral traffic in your analyticsMonthly
ReportingMonthly report and video, plus a mid-month update
Term3 months, then monthly
Code changes on WordPress, Webflow, Shopify, Next.js and React, other custom stacks with repo access.
Ask about a retainerWhat is not included
Everything that sits outside the price, listed here rather than raised later.
Invoiced in USD. The audit is paid up front, retainers monthly in advance, by card or bank transfer.
- Guaranteed mentions, citations or rankings. Nobody can sell those.
- Reddit or Quora posting, review seeding, and link buying.
- Replacing your SEO. This works alongside it.
- Paid ads, social media and PR campaigns.
At least 20 fixes, or your money back
If the audit does not give you at least 20 specific, prioritised fixes, the $750 is refunded in full. That is a promise about the work, which is the kind a vendor can keep. What an AI model says about you afterwards is not something anyone can promise.
What we won’t do
Some of what is sold as AI search optimization is manipulation with a new name. None of it happens here.
No guaranteed mentions or rankings
Nobody controls what ChatGPT, Perplexity, Gemini or Google say. We promise the work itself, and a measurement of what changed.
No Reddit or review seeding
No posting on Reddit or Quora as fake customers, and no buying or seeding reviews. Platforms treat it as manipulation, and it tends to end badly for the brand.
No link buying
No paid links, link schemes or paid brand mentions dressed up as editorial.
No long contracts
3 months, because answers move over weeks. After that it is monthly, with 30 days' notice.
Who is this for, and who isn’t it for?
B2B SaaS and online services
Prospects ask ChatGPT "best X for Y" before they search. You get the list of prompts where competitors appear and you do not, and a fixed price to work on it.
Ecommerce brands, $1-25M
Product recommendations in ChatGPT, Perplexity and AI Overviews skip the results page you paid to rank on. Same method, built around product and category prompts.
Marketing leads with a mandate
The board or the CEO asked you to "do something about AI search". The audit report and 90-day plan are the thing you forward.
A good fit
- B2B SaaS, online services and ecommerce brands selling in the US
- A site with real content and some search traffic already
- WordPress, Webflow, Shopify, Next.js or React, or a custom stack with repo access
- A team that wants the method and the price in writing
Not this
- Local businesses after map-pack rankings. That is local SEO
- A site with no content or traffic yet
- Anyone who wants guaranteed mentions
- Anyone who wants Reddit posts or reviews placed for them
The questions people actually ask
Can you guarantee ChatGPT will recommend us?
No, and nobody can. What an AI model says is decided by the model. What we measure instead is how often your brand appears across an agreed list of prompts on four engines, which sources those engines cite, and how much traffic reaches you from them. Those numbers are in every monthly report.
How long until we see a change?
Technical fixes are read the next time the crawlers visit, which is usually days. The answers themselves typically shift over weeks to months. That is why the minimum term is 3 months: one month is too early to judge.
How do you measure results?
Weekly prompt tracking across 4 engines, reported monthly, and AI referral traffic from chatgpt.com, perplexity.ai and others in your own analytics.
Is this just SEO?
It overlaps. Good SEO helps. The work is measured in prompts instead of keywords, in the sources engines cite instead of backlinks, and the pages are written so a direct answer can be lifted out of them. It runs alongside your SEO rather than replacing it.
Do you post on Reddit for us?
No. Not as you, not as fake customers, and not through anybody else.
Do we need to change CMS?
No. Fixes go into WordPress, Webflow, Shopify, Next.js and React, and other custom stacks where we get repo access.
Who does the work?
Duncan, one engineer, in Glasgow and working US afternoons. The person who audits your site is the person who changes its code and replies to your email. Not an offshore team.
Do you need a call?
No. Everything is written and recorded: the snapshot, the audit report, the walkthrough video and the monthly reports. If you would like a call, one is available.
Which engines do you track?
ChatGPT, Perplexity, Gemini and Google AI Overviews. The free snapshot runs on ChatGPT and Perplexity.
Can we cancel?
Yes, after the first 3 months, with 30 days' notice by email.
Your snapshot or a scope within 1 US business day
The snapshot is free. The audit is $750, and retainers start at $1,500 a month. Tell us what you sell and to whom, and the reply comes back written down.
No call needed. Every price is already on this page.